a.v. bridgewater (1999) an overview of fast pyrolysis of biomass

An overview of fast pyrolysis of biomass

A.V. Bridgwater a,*, D. Meier b, D. Radlein c

aBio-Energy Research Group, Chemical Engineering and Applied Chemistry Department, Aston University, Birmingham B4 7ET, UKbBFH-Institute for Wood Chemistry, Leuschnerstrasse 91, Hamburg D-21031, Germany

cRTI Ltd, 110 Ban Place, Unit 5, Waterloo, Ontario, N2V 1Z7, Canada

Abstract

Biomass fast pyrolysis is of rapidly growing interest in Europe as it is perceived to oer significant logistical andhence economic advantages over other thermal conversion processes. This is because the liquid product can be stored until

required or readily transported to where it can be most eectively utilised. The objective of this paper is to review the designconsiderations faced by the developers of fast pyrolysis, upgrading and utilisation processes in order to successfully imple-ment the technologies. Aspects of design of a fast pyrolysis system include feed drying; particle size; pretreatment; reactor

configuration; heat supply; heat transfer; heating rates; reaction temperature; vapour residence time; secondary cracking;char separation; ash separation; liquids collection. Each of these aspects is reviewed and discussed. A case study shows theapplication of the technology to waste wood and how this approach gives very good control of contaminants. Finally the

problem of spillage is addressed through respirometric tests on bio-oils concluding with a summary of the potential con-tribution that fast pyrolysis can make to global warming.# 1999 Elsevier Science Ltd. All rights reserved.

Keywords: Biomass; Fast pyrolysis; Renewable energy; Fuels; Chemicals; Thermal conversion

1. Introduction

Renewable energy is of growing importance in satis-

fying environmental concerns over fossil fuel usage andits contribution to the Greenhouse Eect. Wood andother forms of biomass are some of the main renewableenergy resources available and provide the only source of

renewable liquid, gaseous and solid fuels. Wood and bio-mass can be used in a variety of ways to provide energy:

. by direct combustion to provide heat for use inheating, for steam production and hence elec-tricity generation;

. by gasification to provide a fuel gas for combus-tion for heat, or in an engine or turbine for elec-tricity generation;

. by fast pyrolysis to provide a liquid fuel that cansubstitute for fuel oil in any static heating orelectricity generation application. The liquid can

also be used to produce a range of speciality andcommodity chemicals

Fast pyrolysis can directly produce a liquid fuel frombiomass which can be readily stored or transported.

2. Fast pyrolysis principles

Biomass is a mixture of hemicellulose, cellulose, ligninand minor amounts of other organics which each pyr-

olyse or degrade at dierent rates and by dierentmechanisms and pathways. Lignin decomposes over awider temperature range compared to cellulose and

hemicellulose which rapidly degrade over narrowertemperature ranges, hence the apparent thermal stabilityof lignin during pyrolysis. The rate and extent of

decomposition of each of these components depends onthe process parameters of reactor (pyrolysis) tempera-ture, biomass heating rate and pressure. The degree ofsecondary reaction (and hence the product yields) of the

gas/vapour products depends on the time-temperaturehistory to which they are subjected before collection,which includes the influence of the reactor configuration.

0146-6380/99/$ - see front matter # 1999 Elsevier Science Ltd. All rights reserved.PI I : S0146-6380(99 )00120-5

Organic Geochemistry 30 (1999) 14791493

www.elsevier.nl/locate/orggeochem

* Corresponding author. Tel.: +44-121-359-3611, ext. 4647.

E-mail address: [email protected] (A.V. Bridgwater).

Although some research has been carried out on the indi-vidual components of biomass, most applied and largerscale work has focused on whole biomass as the cost of pre-separation is considered too high. In addition, the separa-

tion and recovery of pure forms of lignin and hemicelluloseare dicult due to structural changes in their processing,although pure cellulose is relatively easy to produce.

Fast pyrolysis is a high temperature process in whichbiomass is rapidly heated in the absence of oxygen. As aresult it decomposes to generate mostly vapours and

aerosols and some charcoal. Liquid production requiresvery low vapour residence time to minimise secondaryreactions of typically 1 s, although acceptable yields can

be obtained at residence times of up to 5 s if the vapourtemperature is kept below 400C. After cooling andcondensation, a dark brown mobile liquid is formedwhich has a heating value about half that of conven-

tional fuel oil. While it is related to the traditional pyr-olysis processes for making charcoal, fast pyrolysis is anadvanced process which is carefully controlled to give

high yields of liquid. Research has shown that maximumliquid yields are obtained with high heating rates, at reac-tion temperatures around 500C and with short vapourresidence times to minimise secondary reactions. A com-pilation of published data is shown in Fig. 1 for typicalproducts from fast pyrolysis of wood (Toft, 1996). Fast

pyrolysis processes have been developed for production offood flavours (to replace traditional slow pyrolysis pro-cesses which had much lower yields), speciality chemicalsand fuels. These utilise very short vapour residence times

of between 30 and 1500 ms and reactor temperaturesaround 500C. Both residence time and temperature con-trol is important to freeze the intermediates of most

chemical interest in conjunction with moderate gas/vapour phase temperatures of 400500C before recoveryof the product to maximise organic liquid yields.

The essential features of a fast pyrolysis process are:

. very high heating and heat transfer rates, whichusually requires a finely ground biomass feed;

. carefully controlled pyrolysis reaction temperatureof around 500C in the vapour phase, with shortvapour residence times of typically less than 2 s;

. rapid cooling of the pyrolysis vapours to give thebio-oil product.

The main product, bio-oil, is a miscible mixture ofpolar organics (about 7580 wt%) and water (about 2025 wt%). It is obtained in yields of up to 80 wt% in

total (wet basis) on dry feed, together with by-productchar and gas which are, or can be, used within the pro-cess so there are no waste streams.Liquids for use as fuels can be produced with longer

vapour residence times (up to around 5 s) and over awider temperature range although yields might beaected in two ways: secondary volatiles decomposition

at temperatures above 500C and condensation reac-tions at gas/vapour product temperatures below 400C.Most woods give maximum liquid yields of up to 80

wt% dry feed basis at 500520C with vapour residencetimes not more than 1 s. Very short residence timesresult in incomplete depolymerisation of the lignin due

to random bond cleavage and inter-reaction of the lig-nin macromolecule resulting in a less homogenousliquid product, while longer residence times can causesecondary cracking of the primary products, reducing

yield and adversely aecting bio-oil properties. Evidencefrom SEC (selective exclusion chromatography) analysisof the liquids would suggest that the reactor configura-

tion and the dominant mode of heat transfer stronglyinfluences the average molecular weight of the products(McKinley, 1989). This is discussed further below.

Fig. 1. Typical yields of organic liquid, reaction water, gas and

char from fast pyrolysis of wood, wt% on dry feed basis. Fig. 2. Conceptual fluid bed fast pyrolysis process.

1480 A.V. Bridgwater et al. / Organic Geochemistry 30 (1999) 14791493

While a wide range of reactor configurations havebeen operated (Bridgwater, 1999), fluid beds are themost popular configurations due to their ease of opera-tion and ready scale-up. A typical bubbling fluid bed

configuration is depicted in Fig. 2 with utilisation of theby-product gas and char to provide the process heat. Thefigure includes the necessary steps of drying the feed to less

than 10%water tominimise the water in the product liquidoil, and grinding the feed to around 2 mm to give su-ciently small particles to ensure rapid reaction. This con-

figuration is used below for processing waste wood.Several reviews of processes that are advanced tech-

nically and/or commercially available have been pub-

lished (Bridgwater and Evans, 1993; Whiting, 1997). Abrief review of European activities (Bridgwater, 1998)and a comprehensive review of most of the current fastpyrolysis processes has recently been published (Bridg-

water and Peacocke, 1999). Fast pyrolysis of biomassfor liquids began in North America around 1980 andhas seen significant RD&D eort since then with two

commercial organisations oering plants with a perfor-mance guarantee and several demonstration and pilotscale processes.

3. Process characteristics and technology requirements

Although fast pyrolysis of biomass has achievedcommercial status, there are still many aspects of theprocess which are largely empirical and require further

study to improve reliability, performance, product con-sistency, product characteristics and scale-up. This sec-tion summarises these topics.

3.1. Reactor configuration

A variety of reactor configurations have been investi-gated as listed in Table 1. Pyrolysis, perhaps more than

any other conversion technology, has received consider-able creativity and innovation in devising reactor sys-tems that provide the essential ingredients of high

heating rates, moderate temperatures and short vapourproduct residence times for liquids. A thorough reviewof the technologies has recently been completed (Bridg-

water and Peacocke, 1999). There are three main meth-ods of achieving fast pyrolysis.

1. Ablative pyrolysis in which wood is pressedagainst a heated surface and rapidly moved during

which the wood melts at the heated surface andleaves an oil film behind which evaporates. Thisprocess uses larger particles of wood and is typi-

cally limited by the rate of heat supply to thereactor. It leads to compact and intensive reactorsthat do not need a carrier gas, but with the pen-

alty of a surface area controlled system and mov-ing parts at high temperature.

2. Fluid bed and circulating fluid bed pyrolysis whichtransfers heat from a heat source to the biomass

by a mixture of convection and conduction. Theheat transfer limitation is within the particle, thus,requiring very small particles of typically not

more than 3 mm to obtain good liquid yields.Substantial carrier gas is needed for fluidisation ortransport.

3. Vacuum pyrolysis which has slow heating rates butremoves pyrolysis products as rapidly as in the pre-

Table 1

Fast pyrolysis reactors and heating methods

Reactor type Method of heating

Ablative coil Reactor wall heating

Ablative mill Reactor wall (disc) heating

Ablative plate Reactor wall heating

Ablative vortex Reactor wall heating

Circulating fluid bed In-bed gasification of char to heat sand

Cyclone or vortex Reactor wall heating

Entrained flow Char combustion products

Hot sand

Fluid bed Heated recycle gas

Hot inert gas

Partial gasification

Fire tubes

Horizontal bed Fire tubes

Vaccum multiple hearth Hearth heating

Rotating cone Wall and sand heating

Stirred bed Partial gasification of char

Transported bed Recirculated hot sand heated by char combustion

Vacuum moving bed Direct contact with hot surface

A.V. Bridgwater et al. / Organic Geochemistry 30 (1999) 14791493 1481

vious methods which thus simulates fast pyrolysis.Larger particles are needed and the vacuum leads tolarger equipment and higher costs. Total liquidyields are typically lower at up to 6065% compared

to 7580 wt% from the previous two methods.

It is important to remember that pyrolysis always

gives three products gas, liquid and solid of whichthe liquid is a homogenous hydrophilic (oleophobic)mixture of polar organics and water from both the pyr-

olysis reaction and the original water in the feedstock. Asound understanding of the inherent processes will allowany of these products to be maximised and it is the engi-

neers challenge to optimise the process by maximisingproduct quantity and quality while paying proper atten-tion to minimising costs and environmental concerns.

3.2. Heat transfer

There are two important requirements for heat trans-

fer in a pyrolysis reactor:

1. to the reactor heat transfer medium (solid reactor

wall in ablative reactors, gas and solid in fluid andtransport bed reactors, gas in entrained flowreactors);

2. from the heat transfer medium to the pyrolysingbiomass.

Two main ways of heating biomass particles in a fastpyrolysis system can be considered: gassolid heat

transfer as in an entrained flow reactor where heat istransferred from the hot gas to the pyrolysing biomassparticle by primarily convection [for example the Ege-

min (Maniatis et al., 1994), or GTRI (Kovac andONeil, 1989) processes], and solidsolid heat transferwith mostly conductive heat transfer as in ablative pyr-

olysis such as NREL (Diebold and Scahill, 1987) andAston (Peacocke and Bridgwater, 1994). Fluid bed pyr-olysis utilises the inherently good solids mixing totransfer approximately 90% of the heat to the biomass

by solidsolid heat transfer with a probable small con-tribution from gas-solid convective heat transfer of upto 10%. Circulating fluid bed [e.g. CRES (Boukis et al.,

1993)] and transport reactors [e.g. Ensyn (Graham,1991)] also rely on both gassolid convective heat trans-fer from the fluidising gas and solidsolid heat transfer

from the hot fluidising solid although the latter may beless significant than fluid beds due to the lower solidsbulk density. Some radiation eects occur in all reactors.

The important feature of ablative heat transfer is thatthe contact of the biomass and the hot solid abrades theproduct char o the particle exposing fresh biomass forreaction. This removes particle size limitations in certain

ablative reactors (e.g. the NREL vortex reactor), but atthe expense of producing microcarbon which is dicultto remove from the vapour phase and reports to the

liquid product. Attrition of the char from the pyrolysingparticle can also occur in both fluid and circulating fluidbeds, due to contact of the biomass with in-bed solidswhere solids mixing occurs. In fluid bed reactors, how-

ever, attrition of the product char is relatively low and ithas been observed that the char particles have theoriginal particle shape, but are slightly reduced in size

by char layer shrinkage and attrition.Char removal is an essential requirement for large

particles (>2 mm) to avoid slow pyrolysis reactions.

The low thermal conductivity of biomass gives lowheating rates through larger particles which leads toincreased char formation and hot char is known to be

catalytically active. It cracks organic vapours to sec-ondary char, water and gas both during primary vapourformation and in the reactor gas environment. There-fore, its rapid removal from the hot reactor environment

and minimal contact with the pyrolysis vapour productsis essential.Since the thermal conductivity of biomass is very

poor (0.1 W/mK along the grain, ca 0.05 W/mK crossgrain), reliance on gassolid heat transfer means thatbiomass particles have to be very small to fulfil the

requirements of rapid heating to achieve high liquidyields. Claimed temperature increases of 10,000C/smay be achieved at the thin reaction layer but the low

thermal conductivity of wood will prevent such tem-perature gradients throughout the whole particle. Asparticle size increases, liquid yields reduce as secondaryreactions within the particle become increasingly sig-

nificant (Scott and Piskorz, 1984).A consistent method of expressing product yields is

required to remove ambiguities in the comparison of

product yields. It is recommended that the water in thefeed should be discounted in the final pyrolysis productswith only the water of pyrolysis being quoted and the

product yields expressed on a dry feed basis. As a rule ofthumb, the water of pyrolysis is typically 12 wt% of dryfeed.

3.3. Heat supply

The high heat transfer rate that is necessary to heat

the particles suciently quickly imposes a major designrequirement on achieving the high heat fluxes requiredto match the high heating rates and endothermic pyr-

olysis reactions. Reed et al. (1990) originally suggestedthat to achieve true fast pyrolysis conditions, heat fluxesof 50 W/cm2 would be required, but to achieve this in a

commercial process is not practicable or necessary.Each mode of heat transfer imposes certain limita-

tions on the reactor operation and may increase itscomplexity. The two dominant modes of heat transfer in

fast pyrolysis technologies are conductive and convective.Each one can bemaximised or a contribution can bemadefrom both depending on the reactor configuration. The


penalties and interactions are summarised in Table 2below with some speculations on heat transfer modes.For ablative pyrolysis in a vortex reactor, a furnace

arrangement equivalent to an ethylene cracking furnace

has been proposed by the IEA Bioenergy Agreementpyrolysis and liquefaction group (Diebold et al. 1994aand b). Other possibilities to achieve the pyrolysis tem-

peratures and heat transfer rates necessary have inclu-ded vapour condensation such as sodium, inductionheating of the reactor wall and the use of contact elec-

trical heaters. In a circulating fluid bed, the majority ofthe heat transfer will be from the hot circulating sand,typically at a sand to biomass ratio of 20 (Ensyn, 1998),

which therefore requires an ecient sand re-heatingsystem. In a conventional fluid bed the sand requires anexternal heat source which would typically come fromchar in an integrated system.

3.4. Feed preparation

The heat transfer rate requirements described aboveimpose particle size limitations on the feed for somereactors. The cost of size reduction in financial and

energy terms is clear qualitatively but data is not avail-able to define such a penalty associated with the smallparticle sizes demanded of fluid bed and circulating fluid

bed systems. Reactor performance as for example liquidyields is, therefore, not an adequate criterion by itself.

Drying is usually required to less than 10 wt% waterunless a naturally dry material such as straw is available.As moisture is generated in flash pyrolysis, bio-oilalways contains at least about 15% water at an assumed

product yield of around 60 wt% organics and 11 wt%reaction water. This water cannot be removed by con-ventional methods such as distillation. The eect of water

is complex in that it aects stability, viscosity, pH, corro-siveness, and other liquid properties. Selective condensa-tion may reduce the water content of one or more

fractions but at the expense of operating problems and apossible loss of lowmolecular weight volatile components.

3.5. Temperature of reaction

It is necessary to distinguish between temperature ofreaction and reactor temperature. The latter is much

higher due to the need for a temperature gradient toeect heat transfer. For fast pyrolysis the lower limit onwood decomposition is approximately 435C forobtaining acceptable liquid yields of at least 50% withlow reaction times.The eect of temperature is well understood in terms

of total product yield with a maximum at typically 500520C for most forms of woody biomass. Other cropsmay have maxima at dierent temperatures. The eect

of temperature is less well understood in terms of pro-duct fuel quality. Work by the University of Waterloo

Table 2

Reactor types and heat transfer

Reactor type Suggested mode

of heat transfer

Advantages/disadvantages/features

Ablative 95% Conduction

4% Convection

1% Radiation

Accepts large size feedstocks

Very high mechanical char abrasion from biomass

Compact design

Heat supply problematical

Heat transfer gas not required

Particulate transport gas not always required

Circulating fluid

bed

80% Conduction

19% Convection

1% Radiation

High heat transfer rates

High char abrasion from biomass and char erosion

leading to high char in product

Char/solid heat carrier separation required

Solids recycle required; Increased complexity of system

Maximum particle sizes up to 6 mm

Possible liquids cracking by hot solids

Possible catalytic activity from hot char

Greater reactor wear possible

Fluid bed 90% Conduction

9% Convection

1% Radiation

High heat transfer rates

Heat supply to fluidising gas or to bed directly

Limited char abrasion

Very good solids mixing

Particle size limit

has demonstrated the eects of ash, DP (degree ofpolymerisation), heating rate and reactor temperatureon chemical yields (Piskorz et al., 1988). As progress ismade in defining bio-oil quality better in fuel terms, the

secondary gas/vapour phase history may be moreimportant. At prolonged residence times (>1 s), thelignin derived fraction may be further depolymerised to

produce a more homogeneous liquids. This is alsoinflenced by the reactor configuration. Work byMcKinley (1989) has demonstrated that liquid pro-

duced in an ablative pyrolysis reactor has a much lowermolecular weight average due to depolymerisation andcracking of the liquids on the metal surface prior to

vaporisation.

3.6. Vapour residence time

The eect of vapour residence time on organic liquidyield is relatively well understood although the interac-tion of temperature and residence time is less under-

stood. Studies by, for example, Diebold and Scahill(1987), Liden et al. (1988) and Gorton and Knight(1984) have attempted to interlink both primary liquids

formation and secondary cracking, but one essentialcomponent which is neglected is the variation of wateryield with temperature and residence time. There is

extensive literature on the mechanisms and pathways offast pyrolysis which has been reviewed (Peacocke 1994).It is believed that at temperatures below 400C, sec-ondary condensation reactions occur and the average

molecular weight of the liquid product decreases. Bor-oson et al. (1989) have demonstrated that the averagemolecular weight decreases with the degree of secondary

reaction, i.e. increasing residence time and temperature.For chemicals, it is considered necessary to freeze

the process at the appropriate time-temperature point in

the envelope to maximise yield. In one case this has ledto a commercial reactor design where vapour residencetimes down to 30 ms are claimed (Enersearch, 1992).Fuels have less specific process requirements and most

work has focused on maximising liquid yield rather thanproduct quality. The window for fuel productionrequires more R and D to better understand the pro-

cesses and match the product quality requirement toprocess parameters. There is no definition of productquality in terms of physical or chemical properties or

composition, and this area will need to be addressed asmore applications are tested and alternative supplies ofbio-fuel-oil become available.

3.7. Secondary vapour cracking

Long vapour residence times and high temperatures

(>500C) cause secondary cracking of primary productsreducing yields of specific products and organic liquids.Lower temperatures (400C to mini-mise liquid deposition and collection.At present, there are no recognised design methods

and most work has been empirical and specific to thecharacteristics of the feedstock being processed. Com-mercial liquids recovery processes are usuallyproprietary and may be specific to individual feedstocks,

reactor configurations and products.

3.9. Char separation

Some char is inevitably carried over from cyclonesand collects in the liquid. Subsequent separation has

proved dicult. Some success has been achieved with hotgas filtration in a ceramic cloth bag house filter (Diebold etal., 1993) and also candle filters for short run durations.

Liquid filtration has also proved dicult as the liquid canhave a gel-like consistency, apparently due to some inter-action of the lignin-derived fraction with the char.This aspect of char reduction and/or removal will be

increasingly important as more demanding applicationsare introduced which require lower char tolerances interms of particle size and total quantity. Possible solutions


include changing process conditions to reduce the nature ofthe pyrolytic lignin, increasing the degree of depolymerisa-tion of the lignin-derived fraction of the liquid, changingthe feedstock to one with a lower lignin content, or adding

chemicals to the liquid for example to improve handlingproperties or reduce charlignin interactions.It must not be forgotten that an alternative solution is

to modify the application to accept a high char contentbio-fuel-oil.

3.10. Ash separation

The alkali metals from biomass ash are present in the

char in relatively high concentrations and cannot bereadily separated except by hot gas filtration which isundergoing development as reported variously in theseproceedings.

4. Product characteristics

4.1. Product quality

The elemental and chemical composition of pyrolysisoils is very much dependent on the pyrolysis conditionsunder which they are produced. Maximum yields are

obtained at temperatures in the range 450550C andresidence times of 0.55 s depending on the particularprocess; these being typical conditions of fast pyrolysis. Inthis case the oil is highly oxygenated, indeed being not very

dierent in composition from the feedstock (see Table 3).

Typical characteristics of fast pyrolysis liquid aregiven in Table 4. They can vary considerably accordingto the feed material and its characteristics, the pyrolysisprocess parameters and the liquid collection parameters

of which temperature of liquid collection system andmethod of collection are particularly important.The high oxygen content is indicative of the presence

of many highly polar groups leading to high viscositiesand boiling points as well as relatively poor chemicalstability. As the table shows a significant fraction of the

oxygen is present as water. Indeed the description ofpyrolysis liquids as oils is somewhat misleading asthey have low lubricity (Oasmaa, 1997). On account of

their high oxygen content and hydrophilic character it isnot surprising that fast pyrolysis oils are mostly inso-luble in hydrocarbon solvents.Liquid yields decrease at high temperatures and/or

long residence times. The liquid also becomes increas-ingly deoxygenated and at very high temperatures evenpolynuclear aromatic hydrocarbons may form. Fig. 3

shows a widely accepted simplified model of the princi-pal pyrolysis pathways. A medium heating value gas isthe predominant product at high temperatures and high

heating rates. The medium (450550C) temperature

Table 3

Typical elemental and water content of fast pyrolysis bio-oils

Component Wt%

Water 2030Carbon 4447Hydrogen 67Oxygen 4648Nitrogen 00.2

Table 4

Fast pyrolysis liquid characteristics typical data

Moisture content 25%

pH 2.5Specific gravity 1.20Elemental analysis (moisture free basis)C 56.4%H 6.2%N 0.2%S

regime but with high heating rates is characteristic offast pyrolysis and gives mainly condensable vapours andaerosols. Low temperature, slow pyrolysis gives mostlychar. The yields and chemical character of the liquid

product in the various pyrolysis regimes is summarisedqualitatively in Fig. 4.Anaerobic biomass degradation over geologic time

scales is thought to produce mainly peat, lignite andcoal. This is entirely consistent with the model presentedhere as such processes are characterised by low tem-

peratures and very long residence times; conditionswhich minimise liquid and maximise solid products. Onthe other hand petroleum is thought to be derived

mostly from marine organisms. Therefore there isexpected to be very little relationship between petroleumand pyrolysis oils, especially those from fast pyrolysis.Hydrothermal processes for the conversion of ligno-cel-

lulosic biomass to liquids of low oxygen content are

known, for example the PERC (Joyner et al., 1993) andthe HTU (Goudriaan and Peferoen, 1990) processes.There is probably a closer relationship of those pro-cesses to petroleum and bitumen generation. By con-

trast, under fast pyrolysis conditions secondary reactionsare minimised, many functional groups are preservedand only small changes in the overall oxygen content

occur. Indeed, it is possible to identify most individualchemical components as originating from either the lig-nin or cellulosic components of the biomass feedstock.

The complex interaction of pyrolysis time and tem-perature on liquid product quality has not beenexplored, at least partly because the characteristics of

pyrolysis oil for dierent applications have not beendefined and there is no standard pyrolysis liquid. It isthis definition of oil quality that is a major uncertaintyand requires to be defined by potential users and may

dier by application. While there are set standards and

Fig. 4. Qualitative illustration of the yields and character of pyrolysis oils as a function of pyrolysis conditions.


methods of measurement for conventional fuels, analo-gous standards and methods have not yet been definedfor biomass pyrolysis liquids. Density, viscosity, surfacetension and heating value are known to be typical key

properties for combustion applications in boilers, furnacesand engines; but other characteristics such as char leveland particle size and ash content may have a major eect.

4.2. Unusual characteristics of bio-oil

The special features and characteristics of bio-oilrequire consideration in any application including pro-

duction, storage, transport, upgrading and utilisation.These are summarised in Table 5 below. The conclusionis that some problems are soluble and some are moreintractable but none are insoluble. Recognition of pro-

blems and awareness of potential diculties is a majorcontribution to dealing with any of the above diculties.

5. Applications

A summary of the opportunities is shown in Fig. 5.Within Europe the most promising application is seen as

Table 5

Characteristics of bio-oil and methods for modification

Characteristic Eect Solution

Suspended char Erosion

Equipment blockage

Combustion problems from slower

rates of combustion

Sparklers can occur in combustion

leading to potential deposits and

high CO emissions

Hot vapour filtration

Liquid filtration

Modification of the char for

example by size reduction so that

its eect is reduced

Modification of the application

Alkali metals Deposition of solids in combustion

applications including boilers,

engines and turbines

In turbines the damage potential is

considerable particularly in high

performance machines with, for

example, coated blades

Hot vapour filtration

Processing or upgrading of oil

Modification of application

Pretreat feedstock to remove ash

Low pH Corrosion of vessels and pipework Careful materials selection

Stainless steel and some olefin

polymers are acceptable

Incompatibility with polymers Swelling or destruction of sealing

rings and gaskets

Careful materials selection

High temperature sensitivity Liquid decomposition on hot

surfaces leading to decomposition

and blockage

Adhesion of droplets on surfaces

below 400C

Recognition of problem and appropriate

cooling facilities

Avoidance of contact with hot

surfaces above 500C

High viscosity High pressure drops in pipelines

leading to higher cost equipment

and/or possibilities of

leakage or even pipe rupture

Careful low temperature heating,

and/or addition of water, and/or

addition of co-solvents such as

methanol or ethanol

Water content Complex eect on viscosity,

heating value, density, stability,

pH, homogeneity etc.

Recognition of problem

Optimisation with respect to application

In-homogeneity Layering or partial separation of

phases

Filtration problems

Modify or change process

Modify pyrolysis parameters

Change feedstock to low lignin

Additives

Control water content


electricity production due to the anticipated ability touse raw bio-oil as produced in an engine or turbinewithout the need for extensive upgrading as well as theability to de-couple fuel production from electricity

generation with storage and/or transport of the liquidfuel which is not possible for gasifier products andIGCC (Integrated Gasifier Combined Cycle) systems.

The much higher added value of chemicals is, however,increasingly seem as the most interesting short termcommercial opportunity.

5.1. Combustion

Liquid products are easier to handle and transportthan solids and gases and this is important in combustionapplications and retrofitting existing equipment. Exist-ing oil fired burners cannot be fuelled directly with solid

biomass without major reconstruction of the unit, whichmay not be attractive in uncertain fuel markets, howeverbio-oils are likely to require only relatively minor mod-

ifications of the equipment or even none in some cases. Theproblems that have been predicted in handling the fuel withhigh viscosity and suspended char have not been found to

as severe or serious as first thought and many organisa-tions are proceeding with successful developments.

5.2. Power generation with liquids

A key advantage of production of liquids is that fuelproduction can be de-coupled from power generation.

Peak power provision is thus possible with a muchsmaller pyrolysis plant, or liquids can be readily trans-ported to a central power plant using engines or turbine.

There are additional benefits from potentially higherplant availability from the intermediate fuel storage.The economics of power generation suggest that a niche

of up to 10 MWe is available for exploitation.Bio-oil has been successfully fired in a diesel test

engine, where it behaves very similar to diesel in terms

of engine parameters and emissions. Over 400 h operationhave been achieved (Leech, 1997). A diesel pilot fuel isneeded, typically 5% in larger engines, and no sig-nificant problems are foreseen in power generation up to

15 MWe per engine. A 2.5 MWe gas turbine has alsobeen successfully tested although for not many hoursoperation to date (Andrews et al., 1994).

5.3. Chemicals

Several hundred chemical constituents have beenidentified to date, and increasing attention is being paidto recovery of individual compounds or families of che-

micals. The potentially much higher value of specialitychemicals compared to fuels could make recovery ofeven small concentrations viable. An integratedapproach to chemicals and fuels production oers

interesting possibilities for shorter term economicimplementation (Radlein, 1998).Chemicals that have been reported as recovered include

polyphenols for resins with formaldehyde, calcium and/ormagnesium acetate for biodegradable de-icers, fertilisers,levoglucosan, hydroxyacetaldehyde, and a range of fla-

vourings and essences for the food industry. There are sub-stantial problems to be overcome is establishingmarkets forthe less common chemicals and devising low cost and e-

cient separation and refining techniques. The only currentlyviablemarket opportunity is for speciality food flavourings.

6. Fast pyrolysis of contaminated wood

As an example of the application of fast pyrolysis,

some work on fluid bed fast pyrolysis of wastewood isincluded to show how fast pyrolysis can be applied toboth derive a useful liquid fuel product and also fix the

contaminants in the residual char. This work was car-ried out at the Institute of Wood Chemistry in Ham-burg, Germany.

6.1. Wood waste

Wood waste creates an increasing problem in Ger-

many. It is estimated that 68 million tonnes of wastewood arise every year. More restrictive environmentallaws will not allow landfill in the near term and there-

fore the application of new technologies is under inves-tigation. On the other hand, an increasing market forwaste wood already exists. Actual prices for dierent

types of waste wood are listed in Table 6.

6.2. Fast pyrolysis results

Fast pyrolysis of waste wood has been studied in asmall pilot plant with a capacity of 5 kg/h. A schematicdiagram is presented in Fig. 6.Fig. 5. Applications of pyrolysis liquids.


Typical yields of the main products liquid, gas, andchar from untreated wood are depicted in Fig. 7. Oilyields reach almost 70% (wet basis) at 475C reactortemperature. With increasing temperature the gas yields

increase and the char yields decrease correspondingly.Therefore, all subsequent experiments with waste woodwere done at a temperature of 475C. Wood wasimpregnated with inorganic preservatives containingchromium, copper, and boron salts (CCB). This type ofpreservative is amongst creosote the most commonly

used in Germany. After fast pyrolysis, the metal con-tents were measured in the oil and char fraction.The results are presented in Table 7. It is clear that

the heavy metals are collected with the char and that theliquid contains only traces of metals. Real samples froma collection site were also pyrolysed and the results arepresented in Table 7. The real samples show the same

behaviour as those prepared in the laboratory.Other preservatives which are widely used are com-

plexes of alumina and copper (see Fig. 8). The fast

pyrolysis of wood impregnated with these complexesgave similar results: alumina and copper are pre-dominantly found in the char fraction. The oil containsonly traces of these compounds.

6.3. Conclusions

From the results it can be concluded that fast pyr-olysis of waste wood seems to be a viable route fordisposal of contaminated wood waste. Inorganic con-

taminants are almost completely recovered in the charfraction. The oil can be used for energy production or asa source for chemical feedstocks. Further investigation

with chlorinated compounds are ongoing and resultswill be presented later.

7. Biodegradability and fertilizers

7.1. Introduction

Issues of bio-oil toxicity to humans are relevant to itshandling and distribution. However toxicity to micro-

organisms in particular is relevant to its biodegradationin the environment. Some results from respirometricbiodegradation studies are discussed. They lead to the

novel suggestion that bio-oil based fertilisers could playa role in carbon sequestration in soils for the abatementof greenhouse gas emissions.There have been several studies of toxicological eects

of biomass-derived fast pyrolysis bio-oil but they havebeen mostly in regard to the potential hazards asso-ciated with human contact. For example, Gratson

Table 6

Average disposal costs for wood wastes in Germany

Material Cost, euros

Untreated wood 47

Fruit boxes 35

Used furniture 86

Mixed woods 85

Painted wood 96

Demolition wood 68

Impregnated wood 237

Windows with glass 146

Fig. 6. Schematic diagram of pilot plant for fast pyrolysis of waste wood.


(1994) has discussed dermal toxicity, eye irritation,inhalation toxicity and Ames mutagenicity and a reviewof bio-oil toxicity has been carried out by Diebold(1999). On the other hand much less is known about its

fate in the environment and, in particular, the responseof micro-organisms to it. This is important if bio-oilshould emerge as a commodity as that would imply the

shipment of large quantities with the accompanyinghazard of spills. Some of the results from the firstrespirometric study dealing with this issue are discussed

below.

7.2. Fertilisers

A related aspect to this environmental concern is theproposal that bio-oils can be readily converted to nitro-

gen slow-release fertilisers by reaction with nitrogensources like ammonia, urea and proteinacous materialslike manure (Radlein et al., 1997). The favourable

results obtained imply that, at least in this modifiedform, bio-oil is not a hazard to soil micro-organisms.This may prove to be an important observation in the

context of another topical environmental issue, namelythe mitigation of atmospheric carbon dioxide emissions,since such fertilisers could conceivably have a role to

play in the sequestration of carbon in the soil. A pre-liminary assessment of the potential scope of this pro-posal will be made.

7.3. Biodegradability and respirometry

Recently Piskorz and Radlein (1998) have reported

the results of some respirometric biodegradation studiesof bio-oil in both aquatic and soil environments. This isbelieved to be the first application of this technique to

bio-oil. Some of the results from this study and the typeof information that can be obtained are illustratedbelow.

Fig. 9 displays the results from a respirometric test inwhich various quantities of raw Poplar derived bio-oil(RTI Ltd.) were diluted in 500 ml of water and inocu-lated with 5 ml of Returned Activated Sludge (RAS)

from a local wastewater treatment plant. Nutrient defi-ciency of the oil was compensated by the addition ofKNO3 and Na2HPO4 in amounts proportional to the

dissolved organic carbon (DOC) in each sample so as tomaintain a C:N:P ratio of 30:5:1. The progress of bio-degradation was measured by the uptake of O2 over a

period of time.The general pattern is a fast initial oxygen uptake

followed by a long period of negligible biological activ-ity and a final period of fast oxygen uptake suggesting

that bio-oil contains a mixture of both readily andslowly biodegradable substances.Fig. 10 illustrates the eect on the respirometric beha-

viour of neutralizing the bio-oil with lime (Ca(OH)2) to

Fig. 8. Copper-HDO and alumina-HDO as an example for

wood preservatives.

Table 7

Recovery of Cu and Cr from a variety of wood wastes in pilot

plant experiments

Cr (ppm) Cu (ppm)

Char Oil Char Oil

CCB 31561 4 22560 12

CCB + organic preservative 19768 3 8257 16

Cable drums 14103 13 6921 5

Fences 4748 2 1628 7

Railway sleepers 267 2 932 9

Particle boards 527 2 270 9

Window frames 465 2 681 7

Fig. 7. Typical yields of fast pyrolysis experiments with pilot

plant, wt% on dry feed basis.


Fig. 10. Respirometric plots of neutralized poplar bio-oil and

RAS in water (Piskorz and Radlein, 1998).

Fig. 9. Respirometric plots of un-neutralized poplar bio-oil

and RAS in water (Piskorz and Radlein, 1998).

Fig. 11. Respirometric plots of various neutralized bio-oils and

diesel with RAS in water (Piskorz and Radlein, 1998).

Fig. 12. Respirometric plots of various un-neutralized bio-oils

and diesel with RAS in soil (Piskorz and Radlein, 1998).


pH 7 with the other parameters remaining the same. Aclear enhancement of the biodegradation rates is appar-

ent together with a decrease in the length of the periodof inactivity.Fig. 11 is a comparison of 1 ml of each of two dier-

ent neutralized bio-oils, RTI poplar derived and ENELoak derived bio-oils, and a petroleum derived diesel fueloil. The other parameters of the experiment were the

same as before. This demonstrates that the bio-oils havesubstantially higher degradation rates than a typicalhydrocarbon based fuel oil.

As a final illustration Fig. 12 shows the respirometricresults when the same samples were tested in soil with-out either neutralisation, inoculation or nutrient addi-tion in order to simulate natural environmental

conditions. In this case 3 ml of each of the oils wasadded to 10 g of soil representing a substantial loading.From measurement of the solids contents of the

reactors it was possible to estimate specific biodegrada-tion rates using the methodology of Rozich and Gaudy(1992) with the results summarised in Table 8. These

show that bio-oil degrades approximately twice as fastas diesel oil.To summarise, some of the important conclusions of

the study were that:

. Respirometry is a suitable technique for assess-ment of the degradation behaviour of bio-oils innatural environments.

. Pyrolysis bio-oils biodegrade aerobically at a sub-stantially higher rate than hydrocarbon fuels like

diesel oil.. pH neutralisation of bio-oil enhances its biode-

gradability

7.4. Summary

Bio-oil does not appear to be a significant environ-

mental hazard even in raw form, thereby providing amargin of safety in the event of spills. Biodegradabilityis enhanced by neutralisation with bases like lime.

Neutralisation of bio-oil with nitrogenous bases likeammonia, urea or animal manure renders the nitrogeninto a slow-release form useable as a fertiliser. Suchmaterials represent not only a value-added product

from bio-oil but can also play a role in carbon seques-tration in soils. This is especially so if they are used inthe context of energy plantations. Furthermore they

can play a role in alleviation of the problem of nitrogenrun-o which characterises the use of animal manure.

8. Conclusions

Fast pyrolysis is now accepted as a technology forproducing high yields of liquid fuels that can be used inmany applications as direct substitutes for conventionalfuels or as a source of chemicals. There are some inter-

esting challenges to be faced in developing and modify-ing fast pyrolysis technology, in upgrading the liquidsand adapting applications to accept the unusual beha-

viour and characteristics of the liquid product. Higheradded value products than fuels oer the most challen-ging and interesting opportunities including bulk che-

micals such as fertilisers. Pyrolysis also oers potentialfor waste treatment by fixing contaminants in the charwhile also producing a clean liquid fuel.

References

Andrews, R., Patnaik, P.C., Liu, Q., Thamburaj, 1994. Firing

fast pyrolysis oils in turbines. In: Milne, T. (Ed.), Proceed-

ings of the Biomass Pyrolysis Oil Properties and Combustion

Meeting, NREL/CP-430-7215, pp. 383391.

BBSRC,1999. http://www.res.bbsrc.ac.uk/soils/somnet/index.html.

Boroson,M.L., Howard, J.B., Longwell, J.P., Peters, W.A., 1989.

Product yields and kinetics from the vapour phase cracking of

wood pyrolysis tars. AIChE Journal 35 (1), 120128.

Boukis, I., Maniatis, K., Bridgwater, A.V., Kyritsis, S., Flitris,

Y., Vassilatos, V., 1993. Flash pyrolysis of biomass in an air

blown circulating fluidised bed reactor. In: Bridgwater, A.V.

(Ed.), Advances in Thermochemical Biomass Conversion.

Blackie, pp. 4511164.

Bridgwater, A.V., 1998. The opportunities for fast pyrolysis of

biomass. In: Proceedings 10th EC Bioenergy Conference,

Wurtzburg, June 1998, CARMEN, Germany.

Bridgwater, A.V., 1999. Principles and practice of biomass fast

pyrolysis processes for liquids. J. Analytical and Applied

Pyrolysis 51, 322.

Bridgwater, A.V., Evans, G.D., 1993. An assessment of ther-

mochemical conversion systems for processing biomass and

refuse. Report to UK DTI, 254 pp (ETSU B/T1/00207/REP).

Bridgwater, A.V., Peacocke, G.V.C., 1999. Fast pyrolysis pro-

cesses for biomass. In: Sustainable and Renewable Energy

Reviews, Elsevier, pp. 172.

Diebold, J.P., 1999. A review of the toxicity of biomass pyr-

olysis liquids formed at low temperature, In: Bridgwater,

A.V. et al. (Eds.), Fast Pyrolysis: A Handbook. (CPL 1999,

ISBN 1 872691 07 2) pp. 135163.

Table 8

Estimation of bio-kinetic parameters

Sample Medium Specific rate of

bio-degradation

(fraction/day)

Poplar bio-oil (RTI)

Neutralized Water 0.18

Raw Water 0.122

Raw Soil 0.188

Diesel oil (Shell)

Water 0.072

Soil 0.071


Diebold, J.P., Scahill, J., 1987. Production of primary pyrolysis

oils in a vortex reactor. In: Production, Analysis and

Upgrading of Oils from Biomass, ACS Series, Denver, Col-

orado, April, pp. 2128.

Diebold, J.P., Czernik, S., Scahill, J.W., Philips, S.D., Feik,

C.J., 1993. Hot-gas filtration to remove char from pyrolysis

vapours produced in the vortex reactor at NREL. In: Milne,

T.A. (Ed.), Biomass Pyrolysis Oil Properties and Combus-

tion Meeting. NREL-CP-430-7215, pp. 90108.

Diebold, J.P., Beckman, D., Bridgwater, A.V., Elliott, D.C.,

Solantausta, Y., 1994a. IEA technoeconomic analysis of the

thermochemical conversion of biomass to gasoline by the

NREL process. In: Bridgwater, A.V. (Ed.), Advances in Ther-

mochemical Biomass Conversion. Blackie, pp. 13251342.

Diebold, J.P., Elliott, D.C., Bridgwater, A.V., Solantausta, Y.,

Beckman, D., 1994b. Assessment of Liquefaction and Pyr-

olysis Systems, 230 pp, VTT on behalf of the IEA Bioenergy

Agreement, VTT.

Enersearch, Ontario Ministry of Energy, 1992. Project Sum-

mary on Ensyn Technologies RTP Process.

Ensyn, 1998. US Patent 5,792,340, Method and apparatus for a

circulating bed transport fast pyrolysis reactor system, 11

August.

Gorton, C.W., Knight, J.A., 1984. Oil from biomass by

entrained-flow pyrolysis. Biotech. and Bioeng. Symp., No 14,

pp. 1420.

Goudrian, F., Peferoen, D.G.R., 1990. Liquid fuels from bio-

mass via an hydrothermal process. Chem. Eng. Sci. 45, 2729

2734.

Graham, R.G., 1991. Flash-pyrolysis liquid fuels production.

In: Grassi, G. (Ed.), Proceedings 1st EC Forum on Elec-

tricity Production from Biomass and Solid Wastes by

Advanced Technologies. Florence, Italy, 2729 November.

Gratson, D.A., 1994. Results of toxicological testing of whole

wood oils derived from the fast pyrolysis of biomass. In:

Proc. Biomass Pyrolysis Oil Properties ands Combustion

Meeting, Estes Park, Col. USDA, pp. 203211.

Jenkinson, D.S., 1990. The Turnover of Organic Carbon and

Nitrogen in Soil. Phil. Trans. Roy. Soc. London B329, 361

368.

Joyner, H.S., Vaughan, B.M., White, D.H., Wolf, D., 1993.

MSW and biomass to liquid fuels by packaged liquefaction

plants. In: Proc. First Biomass Conference of the Americas,

Vermont. NREL Publication CP-200-5768, 1993, Vol. 2, pp.

964977.

Kovac, R.J., ONeil, D., 1989. The Georgia Tech entrained

flow pyrolysis process. In: Ferraro, G.-L, Maniatis, K., Bue-

kens, A. Bridgwater, A.V. (Eds.), Pyrolysis and Gasification.

Elsevier Applied Science, pp. 169179.

Leech, J., 1997. Running a dual fuel engine on crude pyrolysis

oil. In: Kaltschmitt, M., Bridgwater, A.V. (Eds), Biomass

Gasification and Pyrolysis, CPL, pp. 495497.

Liden, A.G., Berruti, F., Scott, D.S., 1988. A kinetic model for

the production of liquids from the flash pyrolysis of biomass.

Chem. Eng. Comm. 65, 207221.

Maniatis, K., Baeyens, J., Peeters, H., Roggeman, G., 1994.

The Egemin flash pyrolysis process: commissioning and

initial results. In: Bridgwater, A.V. (Ed.), Advances in ther-

mochemical biomass conversion. Blackie, pp. 12571264.

McKinley, J., 1989. Biomass liquefaction: centralised analysis,

final report., DSS File No. 23216-4-6192, Energy, Mines and

Resources Ministry: Ottawa, Canada.

Oasmaa, A., Leppamaki, E., Koponen, P., Levander, J.,

Tapola, E., 1997. Physical Characterisation of Biomass-

based Pyrolysis Liquids. VTT Publications, Espoo.

Peacocke, G.V.C., Bridgwater, A.V., 1994. Ablative plate pyr-

olysis of biomass for liquids. Biomass and Bioenergy 7 (16),

147154.

Peacocke, G.V.C., 1994. Ph.D. thesis, Aston University.

Piskorz, J., Radlein, D., 1998. Determination of biodegrada-

tion rates of bio-oil by respirometry, In: Bridgwater, A.V. et

al. (Eds.), Fast Pyrolysis: A Handbook. (CPL 1999, ISBN 1

872691 07 2) pp. 119134.

Piskorz, J., Radlein, D., Scott, D.S., Czernik, S., 1988. Liquid

products from the fast pyrolysis of wood and cellulose. In:

Bridgwater, A.V., Kuester, J.L. (Eds.), Research in Thermo-

chemical Biomass Conversion. Elsevier Applied Science, pp.

557571.

Radlein, D., 1998. PyNe Newsletter (5) and (6), Aston Uni-

versity, UK.

Radlein, D., Piskorz, J., Majerski., P., 1997. Method of pro-

ducing slow-release nitrogenous organic fertilizer from bio-

mass. US Patent 5,676,727, 1997 and European Patent

Application 0716056.

Reed, T.B., Diebold, J.P., Desrosiers, R., 1990. Perspectives in

heat transfer requirements and mechanisms for fast pyr-

olysis. In: Proceedings of the Specialists Workshop on Fast

Pyrolysis of Biomass, Copper Mountain, Colorado, SERI/

CP-622-1096, pp. 720.

Rozich, A.F., Gaudy Jr., A.F., 1992. Design and operation of

activated sludge process using respirometry. Lewis Publish-

ers, Chelsea, MI, pp. 7796.

Scott, D.S., Piskorz, J., 1984. The Continuous Flash Pyrolysis

of Biomass. Can. J. Chem. Eng. 62 (3), 404412.

Toft, A.J., 1996. Ph.D. thesis, Aston University, Birmingham,

UK.

Whiting, K.J., 1997. The market for pyrolysis and gasification

of waste. Juniper Consultancy Services Ltd, UK.


a.v. bridgewater (1999) an overview of fast pyrolysis of biomass

Documents